Aneurysmal formation of periventricular anastomosis is associated with collateral development of Moyamoya disease and its rupture portends poor prognosis: detailed analysis by multivariate statistical and machine learning approaches.

Abstract

Periventricular anastomosis (PA) is the characteristic collateral network in Moyamoya disease (MMD). However, PA aneurysms are rare, resulting in limited knowledge of their clinical significance. We aimed to elucidate the associated factors and clinical outcomes of PA aneurysms. We reviewed MMD patients who underwent digital subtraction angiography in our institution between December 2001 and March 2023. Genetic analysis was conducted in several cases, and PA aneurysm-positive patients were identified. PA score was defined as the grades of development of periventricular anastomosis. Multivariate analysis and machine learning approaches were used to investigate the significance of the disease and factors associated with PA aneurysm positivity. A total of 301 hemispheres (171 patients) were included. PA aneurysm occurred in 8 hemispheres of MMD (2.7%). PA aneurysm was associated with higher initial modified Rankin scale (mRS) scores (OR, 2.61; 95% CI, 1.45-4.70) and higher PA scores (OR, 1.60; 95% CI, 1.06-2.40). This predisposition was corroborated by gradient boosting and random forest algorithms. Further analysis revealed that PA aneurysm was a risk factor for future hemorrhagic stroke events (HR, 8.29; 95% CI, 1.44-47.7). Among patients in the hemorrhagic-onset group (33 cases), PA aneurysm was a risk factor for worse outcomes (P = 0.008). Principal component analysis also revealed distinct characteristics of hemorrhagic onset aneurysms compared to other MMD cases. Higher PA scores were associated with the presence of aneurysm. PA aneurysm suggests a higher risk of future hemorrhagic strokes, and its rupture portends a worse prognosis.